Autonomic meeting effectiveness and cadence forecasting

Meeting participation data of a meeting of a group of participants in-progress is collected. The meeting participation data is analyzed to identify a topic being discussed in the meeting. Using a trend of affective states of a participant, a future affective state of the participant is forecasted relative to the topic. The future affective state is evaluated to conclude that data contributed by the participant at a future time in the meeting is not likely to progress the topic to completion by at least a specified degree. A cognitive system (cog) trained in the subject-matter is selected. The cog is added to the meeting before the future time and while the meeting is in-progress.